Research On Operation Condition Classification Method For Vibration Monitoring Of Wind Turbine

Data processing method is the core technology in developing condition monitoring system for wind turbines. Previous condition monitoring products usually classify the wind turbine operation condition based on one operation parameter. And a series of monitoring alarm thresholds are defined for each one of the classified wind turbine operation conditions. This leads to a lot of false alarms during the application of condition monitoring system of wind turbine and the problem is urgent to be solved at present. This paper uses multiple operation parameters to classify the complex and non-stationary wind turbine operation conditions related to vibration for a wind turbine. It defines false alarm rate as a rationality evaluation index of operation condition classification methods and then it proposes respectively two multi-parameter classification methods based on rough set (RS) and support vector machine (SVM) to classify the normal operation conditions for a wind turbine. The calculation results of the actual operation data of a wind farm in North China show that:compared with the traditional one-parameter classification method, the operation condition classification method based on RS and SVM can decrease the false alarm rate obviously.